Yes, it turns: experimental evidence of pearl rotation during its formation

Cultured pearls are human creations formed by inserting a nucleus and a small piece of mantle tissue into a living shelled mollusc, usually a pearl oyster. Although many pearl observations intuitively suggest a possible rotation of the nucleated pearl inside the oyster, no experimental demonstration of such a movement has ever been done. This can be explained by the difficulty of observation of such a phenomenon in the tissues of a living animal. To investigate this question of pearl rotation, a magnetometer system was specifically engineered to register magnetic field variations with magnetic sensors from movements of a magnetic nucleus inserted in the pearl oyster. We demonstrated that a continuous movement of the nucleus inside the oyster starts after a minimum of 40 days post-grafting and continues until the pearl harvest. We measured a mean angular speed of 1.27° min−1 calculated for four different oysters. Rotation variability was observed among oysters and may be correlated to pearl shape and defects. Nature's ability to generate so amazingly complex structures like a pearl has delivered one of its secrets.


Introduction
Pearls are the only unique gems produced by a living organism [1]. Their formation process, either by natural means or through human intervention, is a response to an injury to the mantle tissue or a protection against an aggression or an intrusion by a foreign body. The formation of a natural pearl is the result of an accidental transfer of cells of the upper epithelium layer of the mantle into the connective tissue where they multiply to form a closed cyst called a pearl sac [1,2]. Nowadays, pearls are cultured by introducing a nucleus into an animal and biomineral is secreted around it. There are several types of cultured pearls. Non-nucleated cultured pearls are usually mantle-grown in freshwater mussels, whereas nucleated cultured pearls are usually gonad-grown in saltwater oysters [3]. Most nucleated pearl production is from Pinctada species and is the product of grafting and then rearing these oysters in their natural environment. The grafting process takes place following a surgical operation during which the graft, a small piece of mantle tissue, is inserted into the gonad of the receiving oyster together with a shell-based nucleus. Once inserted into the receiving oyster, the external epithelial cells of the graft multiply to form a pearl sac, a single epithelium, around the nucleus [3][4][5][6]. The pearl sac then starts to deposit calcium carbonate polymorph layers onto the nucleus. This is the starting point of the future pearl (figure 1a−h). The nacreous layer is composed of aragonite tablets surrounded by an organic framework (chitin and matrix proteins) [2]. Each mineral tablet thickness is of the order of submicrometres. The assembly of mineral and organic components is guided by enzymatic and by self-assembly mechanisms, which are not well understood. A rearing period of more than 1 year is needed to produce finally a pearl with a sufficiently thick layer of nacre before marketing. The minimum nacre thickness allowed for marketing is 0.8 mm.
The last 5 years have seen a rise in studies of nacre formation conducted in the combined disciplines of biology, physics, chemistry and crystallography with the aim to extend our knowledge about pearl [4,7,8] and shell [7,[9][10][11][12][13] formation. However, our fundamental understanding of nacre formation remains limited. Few studies are available that explain the pearl mineralization process as well as the formation of defects [14]. A remarkable anomaly on the pearl surface is the presence of circles (figure 1), which intuitively suggests a possible rotation of the nucleated pearl inside the pearl sac of the oyster. Rotation of the nucleated pearl during calcium carbonate polymorph layer deposition around the nucleus is also suggested by various observations. These visual examinations can be done at different levels such as the defects on the surface of the pearl, the steps of growth of nacre layer and more exceptionally with the colour (figure 1). Although these observations strongly suggest nucleated pearl rotation, no characterization of such a movement has ever been described. This can be explained by the difficulty of observation of such a phenomenon in the tissues of a living animal. Evidence of pearl rotation is found in both natural and cultured pearls. It is a universal reality that has never been demonstrated and measured. The phenomenon has been mentioned in a few books or local magazines for pearl farmers, written by the early witnesses of pearl culture in Tahiti [15]. They clearly point to rotation as a necessary explanation for the rings and the pear shapes of Tahitian pearls. More recently, Cartwright et al. [14] suggested, based on microscopy observations of pearls from Pteria sterna and Pinctada margaritifera, a general physical framework for pearl rotation. They argued that pearl rotation is a self-organized phenomenon caused and sustained by physical forces from the growth fronts, and that rotating pearls are an example of a natural ratchet. However, there is no accurate characterization of pearl rotation inside the animal.

Experimental grafts
We implemented two experimental grafts with a professional grafter. The first was performed on 12 April 2012 with 15 magnetic nuclei of 6.66 mm diameter (electronic supplementary material, figure S1), 15 'receiving' Pinctada margaritifera pearl oysters coming from Takarao Island (Tuamotu, French Polynesia) and one 'donor' oyster. After grafting, the 'receiving' oysters were individually maintained in a pocket of a plastic net, as done by pearl farmers, and hung up on a long line in the experimental oyster growing field of Ifremer in the lagoon of Vairao (Tahiti Island). Nucleus rejection was checked every week post-grafting and only receiving oysters that retained their nucleus were kept. After four weeks, oysters were transferred to the oysters' nursery of Ifremer in raceways with continuous flow of seawater and food (microalgae Isochrysis galbana). The first tests on the monitoring system were done 20 days post-grafting.  from the Ifremer growing field were grafted using four different 'donor' oysters and the same nucleus size. After grafting, the 'receiving' oysters were individually maintained in the lagoon of Vairao as was done for the first graft. After 12 days, oysters were transferred to the oysters' nursery of Ifremer and the first data acquisitions started at a post-grafting time of day + 13.

Data acquisitions experiment
The so-called 'magnetometer' system (electronic supplementary material, figures S2 and S3) was specifically engineered for our study to register magnetic field variations with magnetic sensors from movements of a magnetic homemade nucleus (electronic supplementary material, figure S1) inserted in

Mathematical treatment of the data
Vectors associated with magnet positions over time are deduced from the magnetic flux density measured at each sensor using (electronic supplementary material, figure S3) where Pi is the position of the magnet at time i in three-dimensional Cartesian coordinates (x i , y i , z i ); j, sensor number, from 1 to 25; a ij , magnetic flux density (tesla) from the sensor j at time i; P j , Cartesian coordinates for the sensor j on the dome. The time series obtained were linearly filtered using a centred moving average to remove white noise. Durbin-Watson test was computed to test absence of autocorrelation in residuals (α = 0.05) and to adapt the size of the window width used while filtering. The radius (r) and the coordinates of the centre O (a, b, c) of the best-fit sphere to the set of data points were determined using least-squares estimation from sphere equation: with ρ i = x 2 + y 2 + z 2 and n corresponding to the number of values i. The sphere centre O was then shifted at the origin. A central projection of vectors onto the sphere surface was applied for correcting remaining error measurements: with |R i | corresponding to the length of each vector OP i : |R i | = x 2 + y 2 + z 2 . Eventual movement of the nucleus during the recording period was thus graphically represented by a set of points lying on a unit-sphere surface.
Angles Ψ i between successive vectors were then calculated assuming absence of nucleus translation. The axis of rotation was dynamically found every minute using a centred windows size of 130 min, from plane equation using least squares. Thereafter, points were orthogonally projected on the perpendicular  figure S6). The angle Ψ 1 between vectors OP1 and OP2 with respective coordinates (x 1 , y 1 , z 1 ) and (x 2 , y 2 , z 2 ) was thus given by Ψ 1 = arccos (x 1 .x 2 + y 1 .y 2 + z 1 .z 2 ). Angular speed was then deduced by averaging angles Ψ over time of recording. Angular velocities were estimated only for grafted nuclei rotating around axes changing slightly over time.

Experimental validation of the measurement system
To validate the magnetometer system, we analysed and characterized the artificial movements of the magnet. We used a clock mechanism to get a constant rotation with a known speed. The clock system was put up to the monitoring dome with a stick glued to the minute hand mechanism (electronic supplementary material, figure S7A). The stick was long enough to place a magnet (the same as the ones implanted in the grafted nucleus) in the monitoring area. The rotation axis was the stick axis with a speed of 0.5 • min −1 set by the hour hand rotation. The first set of experiments was dedicated to data collection and graphical representations of rotations with different rotational axis of the magnet. The stick axis was fixed with an angle varying from 0 • to 90 • to the horizontal (total recording duration of 8995 min). The magnet was either perpendicular to the stick or once fixed with an angle of 45 • (slanted magnet). The magnetic flux of a motionless magnet horizontally positioned was also measured four times to understand better the magnetometer precision in absence of movement (total duration of 5126 min).
The three-dimensional plots of rotating and motionless magnets are presented in the electronic supplementary material, figure S7B-C. Results of angular speed calculation and radius of the detected spheres are presented in the electronic supplementary material, table S1, for the rotating magnets. Shapes of unfiltered points representing the movements of the rotating magnets over time were laid on a sphere surface. After filtering, points were aligned and represented well the rotation on the hour hand (electronic supplementary material, figure S7B). Conversely, unfiltered points describing positions of the motionless magnet over time formed clouds without particular pattern (electronic supplementary material, figure S7C). In this case, the best-fit spheres for the four recordings were always small (mean radius of 0.15 ± 0.13, with a maximum size at 0.36) in comparison with rotating magnets (mean radius of 4.75 ± 2.31, electronic supplementary material, table S1) and pearls (mean radius of 9.26 ± 1.74, with a minimum at 6.31, N = 15). Points were close together and small position variations were only due to noise. Motionless nucleus could thus be detected using size of the best-fit sphere radius and shape of points. For rotating magnet, angular speed is well approached by the calculations (electronic supplementary material, table S1).

Results and discussion
To investigate the characterization of pearl rotation, the 'magnetometer' system (figure 2 and electronic supplementary material, figures S1, S2, S3 and S4) was specifically engineered to register magnetic field variations with magnetic sensors from movements of a magnetic homemade nucleus inserted in the gonad of a pearl oyster (electronic supplementary material, figure S1). Two oysters were placed under the 'magnetometer' system from the first to the 50th day after the grafting. This time window corresponds to the period necessary for the establishment of the pearl sac [14]. For each measurement, the movements of the nucleus were recorded; these becoming more and more stable with increasing time (figure 3). Aside from 180 • discontinuous rotation of the nucleus observed for the first 12 days, records showed chaotic movements of the nucleus punctuated with motionless period. After these periods of around 30-40 days, the nucleus rotation of the two oysters changed and became continuous for all the recording period (figure 3b−d). All measurements performed on nine oysters showed continuous movements of the nucleus after a minimum of 40 days post-grafting. Some rotations could be rather stable (electronic supplementary material, figure S5A), whereas others could show important variations of the rotation axis position over time (electronic supplementary material, figure S5B). Angular velocities are also variable among records. We obtained a mean angular speed of 1.27 • min −1 calculated for four different oysters, with a maximum mean per recording period of 2.10 ± 0.82 • min −1 (s.d.) and a minimum of 0.40 ± 0.15 • min −1 (s.d.). Under assumption of rotations around a fixed axis, these mean velocities correspond to, respectively, one turn in 4 h and 43 min for the mean calculated on four different oysters and 2 h and 51 min and 15 h for extremes.
Rotation variability may be correlated to pearl shape and defects. The last records, on nucleus movement before harvesting, revealed three kinds of pearls for three contrasted rotations ( figure 4).   Rotations around an axis quite stable in time without additional movements inside the pearl sac were associated with a slightly circled pearl with an axis of rotational symmetry (figure 4a−d). Rotations in a preferential zone due to slight displacements of the nucleus in the pearl sac between rotations led to a semi-round pearl with one coloured circle on the top (figure 4b−e). Irregular rotations, with a variable axis position over time and different senses of rotation, gave a non-circled round pearl with consistent colour (figure 4c−f ). According to the hypothesis of Cartwright et al. [14] concerning rotation induced by pearl growth, steps of round pearls would not be necessarily randomly oriented across its surface but could globally share a direction varying over time. Rotation in an irregular way could favour homogeneity of deposits on nucleus creating a round pearl. Here, we experimentally showed that the movement of a round pearl is in several successive directions (figure 4c−f ), thereby validating the hypothesis of Cartwright et al. [14] which said that for a spherical bead, the movement would be random, while for a pearl with rings (figure 1c-f ), movements would be with a single rotation axis (figure 4a−d). This is consistent with our observations. Our results are in accordance with the work of Cartwright et al. [14] bringing the experimental evidence of the theoretical model.
At the end of 40 days post-grafting, the rotation of the pearl appears to be continuous throughout its complete formation, which lasts 12-18 months. This time window corresponds to the period necessary for the establishment of the pearl sac [14]. The secretion of the pearl sac, which is a single epithelium, is totally dependent on the physiology of the animal. These cells, programmed to secrete the components of the shell, begin to deposit material on the pearl nucleus very soon after the graft. The secretion mechanism, caused and sustained by physical forces from the growth fronts, reproduces the structure of the shell in the pearl and was proposed to be responsible for the pearl rotation [14]. As long as the pearl remains in the animal, the pearl sac continues to secrete nacre-forming compounds; nacre lamellae progressively cover the nucleus, so the pearl grows, turns and the pearl sac expands with it.

Conclusion
Cultured pearls are human creations formed by inserting a nucleus and a small piece of mantle tissue into a living shelled mollusc, usually a pearl oyster. Then, the biomineral is secreted by the external epithelial cells of the mantle tissue around the nucleus to produce the pearl ( figure 1). We have demonstrated here that, during this process, the pearl rotates continuously inside the oyster throughout its formation and we have quantified its movement. We showed that a continuous movement of the nucleus inside the oyster starts after a minimum of 40 days post-grafting and continues until the pearl harvest 12-18 months later. Nature's ability to generate so amazingly complex structures like a pearl is impressive and still partially unknown. One of its secrets has been delivered through this study. Such a challenge will strongly encourage the community to continue to develop approaches and knowledge to decipher the biophysical and molecular details of biomineralization processes. Competing interests. We declare we have no competing interests. Funding. Support was provided by Ifremer, Université de la Polynésie française and Direction des Ressources Marines et Minières de la Polynésie française.